Annual Report #1Project No. G-35W02

Seismic Surface Wave Tomography of Waste Sites

Principal Investigator,Dr. Leland Timothy LongProfessor of Geophysics

Research Associates,Dr. Argun KocaogluFrank Williamson

Prepared for:U. S. Department of EnergyEnvironmental Management Science ProgramDE-FG07-96ER14706

October 14, 1997

GEORGIA INSTITUTE OF TECHNOLOGYSchool of Earth and Atmospheric Sciences21 Bobby Dodd Way

Atlanta, Georgia 30332-0340

Executive Summary

for

Seismic Surface Wave Tomography of Waste Sites

The objective of the Seismic Surface Wave Tomography of Waste Sites is todevelop a robust technique for field acquisition and analysis of surface wave data for theinterpretation of shallow structures, such as those associated with the burial of wastes.The analysis technique is to be developed and tested on an existing set of seismic datacovering the K-901 burial site at the East Tennessee Technology Park. Also, a portableprototype for a field acquisition system will be designed and developed to obtainadditional data for analysis and testing of the technique.

The K-901 data have been examined and a preliminary Single ValuedDecomposition inversion has been obtained. The preliminary data indicates a need foradditional seismic data to ground-truth the inversion. The originally proposed gravitydata acquisition has been dropped because sufficient gravity data are now available for apreliminary analysis and because the seismic data are considered more critical to theinterpretation. The proposed prototype for the portable acquisition and analysis systemwas developed during the first year and will be used in part of the acquisition ofadditional seismic data.

Index

Executive Summary

index

Research Goals

Summary of accomplishments

APPENDICES

Appendix I Development Computer Setup and Software

Appendix II Data Analysis

Appendix III Seismic Data Acquisition System, System Description

The design and construction of the field acquisition system are described in themanual for the instrument which is attached as Appendix III. The hard copy of theoperating program referenced in Appendix III has been omitted to save space. Theprograms may be obtained in electronic form by sending an email totim.long@eas.gatech.edu. This manual was prepared by Mr. Frank Williamson.

Appendix IV Validation of Shallow Seismic Structure at the K-901 Burial Grounds, EastTennessee Technology Park.

Research Goals

The objective of this research is to develop a robust method to determine the near-surface shear-wave velocity structure by using measurements of surface-wave groupvelocity. The principal goal is to adapt the surface-wave inversion to a portable computerwhich will allow the image to be developed in the field from successive data samples.We will use previously obtained data from the K-901 site at Oak Ridge NationalLaboratory for developing and testing the analysis method.

The two specific tasks proposed in this research are:

Task 1: The first objective is to perform surface-wave dispersion tomography on existingdata and to evaluate the advantages and disadvantages of various data reduction andinversion methods. The data from the K-901 site at Oak Ridge National Laboratory willbe used. The velocity structure will be obtained from an inversion of surface wave groupvelocity and from direct measurement of velocity using other seismic and geophysicaltechniques.

Task 2: Design and build a portable data acquisition and analysis system that will allowreal-time image development in the field.

Proposed Goals for first year:

Task 1: (at Georgia Tech with advice from ORNL) The analysis for the first year will bethe processing and tomographic inversion of an existing data set for the K-901 BurialGrounds at the East Tennessee Technology Park (formerly known as the Oak Ridge K-25Plant).

a) Examine the data and establish computational compatibility with programs atGeorgia Tech.

b) Develop the software to automatically detect and compute group velocitiesusing the data from the K-901 site

c) Modify the software from the preliminary tests for use in areas with arbitraryboundaries.

d) Compute a singular value decomposition solution for the tomographicinversion.

e) Program and develop the algebraic reconstruction inversion technique for fielddata.

f) Design the acquisition systemg) Find sites to test the system.

Task 2: (at ORNL) Use microgravity data to ground-truth the structure under the K-901sitea) Increase density of stations over the K-901 site.

b) Use inverse methods to model the presumed structure.

Proposed Goals for second year:

A system will be designed around a portable computer with digital acquisitioncapabilities. The system will be designed for use in areas of dimensions of 50 to 100meters and to allow use of simple portable impulsive sources such as mechanical weightdrops or hammers.

Proposed Goals for third year:

The third year will be used for field testing of the portable system and evaluation ofnumerical techniques in the analysis. The techniques will be tested in two other testareas.

Summary of accomplishments during 1997 and proposed goals for 1998 and 1999.

Task 1: (at Georgia Tech with advice form ORNL) The analysis for the first year will bethe processing and tomographic inversion of the existing data set for he Oak Ridge K-25Plant at the Oak Ridge Reservation K-901 site.

a) Examine the data and establish computational compatibility with programs atGeorgia Tech. Task 1 a is complete. The Computer system for thedevelopment of the inversion program and analysis of the K-901 data set isdescribed in Appendix I. The computer is a 200MHx pentium runningLINIX. The analysis of the K-901 data was largely performed using SU,seismic Unix, a non-proprietary seismic analysis program. Other programsused in the analysis include MATLAB, FORTRAN, and VISUAL BASIC.Although the original copy of the K-901 data were not readable, we were ableto convert the original data to a version of SEGY which is compatible withour programs. We were also able to supply header data giving shot andreceiver locations for our analysis.

b) Develop the software to automatically detect and compute group velocitiesusing the data from the K-901 site. Task lb was partially completed. Thegroup velocity travel-time picks as originally computed were neither robustnor sufficiently precise to use directly in an inversion. In order to eliminatenoise, we are developing a method that uses phase coherence as a constraint inpicking group velocities. An arrival time picking routine was implementedand tested for P-wave arrivals (See Appendix II).

For 1998 and 1999, we will continue to examine the problem ofdetermining coherent group velocity arrival times. The difficulty in

pinning down the group velocity is expected to be the major challenge inthe inversion of group velocities.

c) Modify the software from the preliminary tests for use in areas with arbitraryboundaries. Task lc is essentially complete and is not a significantconsideration. The boundary to the K-901 site was irregular and for analysis,boundary data relative to shot and receiver locations had to be incorporated inthe analysis programs.

For 1998 and 1999 arbitrary boundaries will be part of all programs.

d) Compute a singular value decomposition solution for the tomographicinversion. Task Id was started and preliminary interpretations are available.The SVD solution was obtained for surface waves from 10 to 24 Hz, as wellas for the refracted P-wave arrivals. The P-wave data indicated a deep-seated(50 to 100m) refractor of high velocity, probably unweathered limestone. Thepreliminary surface-wave group velocity SVD inversions were contaminatedby scatter in arrival time picks. These preliminary results are shown inAppendix II. The K-901 site data were obtained with 8 Hz geophones. Thefrequencies below 8 Hz are strongly attenuated in such recording instrumentsand are difficult to analyze. In particular, group velocities can have multipleanswers for a given frequency. Consequently, without a record of the low-frequency energy, we found it difficult to identify the portion of the dispersioncurve responsible for the seismogram. In particular, it was difficult todetermine if the reverse dispersion observed in the frequencies above 8 Hzwas caused by a low velocity layer or caused by observing only thefrequencies above the group velocity minimum. We are proposing seismicfield tests to resolve this ambiguity. In either model, synthetic seismogramscan be made to match the observed data for the higher frequencies.

e) Program and develop the algebraic reconstruction inversion technique forfield data. Task le was not attempted during the first year.

For 1998 fall and winter quarter we plan to implement a genericalgebraic reconstruction inversion technique on a portable analysis system.

f) Design the acquisition system Task If was completed. In addition, theproposed goal of building this system during the second year was completedduring the first year. We used an existing computer and digital acquisitionboard to construct a field acquisition system. The design and data recordingprograms for this system are described in Appendix III.

For 1998 we will convert the programs on the acquisition system torun in visual basic under windows NT, in order to allow simultaneous dataacquisition and analysis and to avoid problems with the Dynamicallocation of IRQ in windows 95 and NT. The current program runs underquick basic in DOS 6.2. For 1999 we will consider transfer of the systemto a faster portable computer, depending on the success of field acquisitionand analysis during 1998.

g) Find sites to test system. Task lg is open. We have designated 4 sites withdiffering surface velocity characteristics for the test. One is the Cobb Countytest site, a site for which we have preliminary data and knowledge of thelocation of various buried objects. The other three are the sites of threeearthquake monitors, where the near-surface velocity structure could assistother studies. Other sites will be tested as appropriate and as made availablewhen sufficient auxiliary data are available to make the study useful.

Task 2: (at OWL) Use microgravity data to ground-truth the structure under the K-901site. A) Increase density of stations over the K-901 site. B) Use inverse methodsto model the presumed structure. Task 2 was significantly changed. It wasdetermined that sufficient gravity data were available for a preliminaryinterpretation and that instead of additional gravity data, supplemental seismicvelocity information were more essential for a proper interpretation of the area.It was discovered in the preliminary analysis that proper interpretation of the P-wave arrivals requires a full length refraction line. Also, without a full lengthrefraction line, the development of the surface waves could not be observed.Therefore, we are proposing to obtain refraction line data over the test site. Inorder to resolve the existence of the refractor at depth, a high-frequency reflectionand up-hole velocity surveys are proposed. In order to resolve the fulldevelopment of the surface waves, a refraction line using 1.0 Hz sensors isproposed. Also, an anomalous high velocity shallow structure was found and twoor more fan arrays with 1.0 Hz sensors are proposed to isolate these structures.Appendix IV is a summary of the proposed field measurements and analysis.

For 1998 we plan to obtain the supplemental velocity data. The contractto transfer funds to ORNL (as proposed in the original budget) is currently undernegotiation. If an agreement on a funding method is successful, we hope tocomplete the field work before the end of November and before the weathermakes the work difficult.

APPENDIX I, Development of Computer Setup and Software

For the analysis of the data set collected at the K-901 test site, a 200 MHz Pentium PC is set upas a workstation running LINUX (a full featured Unix-like operating system for x86 computers).The LINUX OS comes with GNU C,C++ and FORTRAN compilers, X Window System, TeXtypesetting package, editors, and many other Unix utilities. Additional software packages installedon the new system includes SU (Seismic Unix) Release 30, SAC (Seismic Analysis Code) 2000,MATLAB 5.0, Pgplot 5.2 (a FORTRAN Graphics Library), XForms (a graphical user interfacetoolkit for X Window systems).

Seismic Unix

The CWP/SU package is a free software package created at the Center for Wave Phenomena. Col-orado School of Mines. The package is a seismic processing environment for UNIX-based machines.The package contains tools for: reading/writing tapes in the SEG-Y format, manipulating seismicdata in the SEG-Y format, Fourier transforms, filtering, synthetic data generation, seismic migra-tion, NMO, DM0, Wavelet transforms, Delaunay triangulation, Postscript Graphics and X-windowsGraphics (Cohen and Stockwell, 1997). All SU programs provide information about themselves whenthe name of the program is typed without any arguments. For example,

% suximage

SUXTMAGE - X-windows IMAGE plot of a segy data set

suximage <stdin [optional parameters] I . . .

Optional parameters :

n2=tr.ntr or number of traces in the data set (ntr is an alias for n2)

dl=tr.d or tr.dt/10A6 sampling interval in the fast dimension=.004 for seismic (if not set)= 1.0 for nonseismic (if not set)

d2=tr.d2 sampling interval in the slow dimension=1.0 (if not set)

fl=tr.fl or tr.delrt/10A3 or 0.0 first sample in the fast dimension

f2=tr.f2 or tr.tracr or tr.tracl first sample in the slow dimension= 1.0 for seismic (if not set)=d2 for nonseismic (if not set)

verbose=0 =1 to print some useful information

tmpdir= if non-empty, use the value as a directory path.prefix for storing temporary files; else if thethe CWP_TMPDIR environment variable is set useits value for the path; else use tmpfile()

Note that for seismic time domain data, the "fast dimension" istime and the "slow dimension" is usually trace number or range.

Also note that "foreign" data tapes may have something unexpectedin the d2,f2 fields, use segyclean to clear these if you can affordthe processing time or use d2= f2= to override the header values ifnot.

See the ximage selfdoc for the remaining parameters.

On NeXT: suximage < infile [optional parameters] | open

suhelp lists the programs available in the distribution. The first two pages of the output is asfollows:

CWP PROGRAMS: (no self-documentation)ctrlstrip feat maxintsdownfort isatty pause upfort

PAR PROGRAMS: (programs with self-documentation)a2bb2adzdvfarithfarithlftnstriph2b

kaperturemakevelmkparfileprplotrayt2drecastregrid3

resampsmooth2smooth3dsmoothint2subsetswapbytestmb

trampunif2unisamunisam2velconvvelpertvtlvz

wkbjxy2zz2xyz

press return key to continue

SU PROGRAMS: (self-documented programs for SU dataSU data is "SEGY" data run through "segyclean"

bhedtopardtltosusegdreadsegycleansegyhdrssegyreadsegywritesetbhedsu3dchartsuabshwsuacorsuaddheadsuaddnoisesuampsuasciisuasciiallsuattributessubfiltsuchartsuchwsuconv

sudivcorsudiv stacksudmofksudmotxsudmovzsueditsufdmod2sufftsufiltersuflipsufracsugaborsugainsugazmigsugetsugethwsuharlansuhilbsuifftsuilogsuimp2d

suklk2filtersukdmig2dsukdsyn2dsukfiltersukfracsukillsulogsumaxsumediansumigpssumigtksumigtopo2dsumixsumutesunmosunullsuoldtonewsuopsuop2supacklsupack2

supickampsuplanesuputsuquantilesuradonsurampsurangesurecipsureducesureflpsvshsurelansuresampsuresstatsushapesushwsusortsuspecfksuspecfxsuspecklk2suspikesustack

sustnpsuswapbytessusynczsusynlvsusynlvcwsusynvxzsusynvxzessutabsutapersutaupsutsqsuttozsutvbandsuunpacklsuunpack2suvcatsuvelansuvibrosuvlengthsuwindsuxcor

sudatumk2dr suimp3dsudatumk2ds suinterpsudipfilt suintvel

supastesupef

supgc

sustat icsustkvelsustolt

suxeditsuzero

press return key to continue

Other utilities providing information on SU programs are suname, sudoc and sufind The SUUser's Manual is a good starting point to learn more about the package and its use in seismic dataprocessing and visualization.

traceview

The shell program traceview can be executed to view a certain trace in a given record section.Below is the contents of the file traceview

#! / b i n / s hfname=$HOME/ornl/data/su/1001- 1099.sushot=$ltrace=$2suwind < $fname key=fldr min=$shot max=$shotl \suwind key=tracf min=$trace max=$tracesuxgraph style=normal

\

Once the above lines are saved in a file called traceview it can be executed. The simplest way toexecute a UNIX shell is to make it executable. This is done by using the chmod command:

% chmod +x traceview

The SU script traceview takes two argument in the command line. The first one is the field recordnumber (shot number) and the second one is the trace number within field record. As an examplelet us view a trace in K-901 data set:

% traceview 1002 44

Here 1002 is the shot point located on the northeast corner and the 44 is the recording point on thenorthwest corner. Figure 1 shows the output of the command.

6 0.5 i 1.5 2 2.5 3 3.5 i

Figure 1: Sample trace viewed by shell script traceview.

shotview

This shell is written to allow display of a given field record.

#! /b in /shfname=$HOME/ornl/data/su/1001 -1099.sushot=$2segyread endian=0 tape=$fnamel \suwind itmax=4095 key=fldr min=$shot max=$shotl \suxwigb style=normal

All 48 channels of the shot gather 1030 is displayed on the screen with the following command:

shotview 1030

The output is shown in Figure /refshotview.

•gill

Figure 2: Sample shot gather viewed by shell script shotview.

traceselect

traceselect extracts a trace out of a record section. The output is an ASCII file containing theheader information followed by the data values. The shell script is as follows:

#! /bin/shfname=$HOME/ornl/data/su/1001-1099.sushot=$ltrace=$2outfile = shot$shot-trace$trace.datsuwind <$fname key=fldr min=$shot max=$shotl \suwind key=tracf min=$trace max=$trace | \suasciiall > $outfile

Then,

% traceselect 1002 44

extracts the 44th trace of shot 1002. The output file is named automatically by the script asshotl002-trace44.dat, and its first 30 lines are

tracl=0 tracr=0 fldr=1002 tracf=44 ep=O cdp=Ocdpt=0 trid=l nvs=8 nhs=0 duse=0 offset=0gelev=0 selev=0 sdepth=O gdel=0 sdel=O swdep=0gwdep=0 scalel=0 scalco=-100 sx=1137804 sy=4247375 gx=1112OO3gy=4282544 counit=0 wevel=0 swevel=0 sut=O gut=Osstat=O gstat=O tstat=O laga=0 lagb=0 delrt=Omuts=0 mute=0 ns=4096 dt=1000 gain=0 igc=Oigi=0 corr=0 sfs=O sfe=O slen=0 styp=0stas = 0 stae = 0 tatyp = O afilf=O afils = 0 nofilf=60nofils=0 lcf=0 hcf=250 lcs=0 hcs=18 year=Oday=0 hour=0 minute=0 sec=0 timbas=O trwf=0grnors=45 grnofr=45 grnlof=0 gaps=O otrav=0 dl=0.000000fl=0.000000 d2=0.000000 f2=0.000000 ungpow=0.000000 unscale=0.000000 ntr=Omark=0

-9.1100e+02-1.1280e+03-1.2180e+03-1.2300e+03-1.0350e+03-6.4100e+02-1.7100e+02

1.4900e+023.1000e+01

-3.3700e+02-4.6100e+02-4.0600e+02-5.2600e+02-6.8500e+02-6.7000e+02-5.2900e+02-1.1000e+02

Reflectivity Program

The name of the program that generates the reflectivity seismograms is reflect2. The programresides in /usr/local/bin so that it is accessible by all users. The program reflect2 rrequires oneinput file which includes the names of the actual input parameters file and two output files. If, forexample, reflect.dat is the name of the input file, it includes three lines each of which indicates filenames:

elastic .inpelastic.psvelastic.sh

Here elastic.inp is the name of the input parameters file. And the last two are names of the outputfiles for PSV and SH motion seismograms. A sample input parameters file is like this:

INPUT-EXAMPLE POINT SOURCE, INHOM:0 0 0 0 1 0 0 1 1 1 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0

1 0 1 10.0000 0.00000.0000 6.0000 200. 200.

20.0000 6.0000 200.0000 200.20.0000 8.0000 200.0000 200.

10.0000 100.0000 10.00006 . 0 0 0 0 0 . 06.0500 6.20000.0000 1.00000.0200 512 0 2 0.2000 0.0000

Once the input parameters file is prepared, the reflectivity program can be executed as follows:

% reflect2 < reflect .dat

The file reflect2.doc located /usr/local/reflect2 describes each of the input fields in the file.This contents of this file are

INPUT DATA (UNIT 98):LINE 1: IDENTIFICATION DENTIF

FORMAT: 10A8LINE 2: SWITCHES ISS(25)

FORMAT: 5(512,2X)

LIST OF SWITCHES

N ISS(N) EXPLANATION

1 =0 FULL WAVE FIELD SEISMOGRAMS= 1 ONLY PP-REFLEXION "=2 ONLY SS-REFLEXION "=3 ONLY PP- AND SS-REFLEXION SEISMOGRAMS

2 =0 TRANSMISSION WITH CONVERSION= 1 TRANSMISSION WITHOUT CONVERSION (TPS=TSP=O.)

168.

.00000000

0

2 0 .10.0000

?

00

400.0000.2000

3 =0 SEISMOGRAM FOR DISPLACEMENT=1 SEISMOGRAM FOR VELOCITY=2 SEISMOGRAM FOR ACCELERATION

4 =0 COMPUTE VERTICAL AND RADIAL COMPONENT= 1 COMPUTE ONLY VERTICAL COMPONENT=2 COMPUTE ONLY RADIAL COMPONENT=3 NO VERTICAL AND RADIAL COMPONENT

5 =1 NO TRANSVERSE (SH) COMPONENT IS COMPUTED=0 TRANSVERSE COMPONENT IS COMPUTED

6 = 1 WRITE OUT ON FILES 21, 22, 23 THE DISPLACE-MENT^, OMEGA) FOR SPECTRAL CONSIDERATIONS

7 =1 NO UPGOING WAVES FROM THE SOURCE ARECONSIDERED

8 =0 LAYER PARAMETER READ IN= 1 FNHOMOGENEOUS VELOCITY-DEPTH-DISTRIBUTION

READ IN (SUBROUTINE INHOM)

9 =0 LAYER PARAMETER T H I C K N E S S ,P-,S-VELOCITY,DENSITY (ISS(8) MUST=01

= 1 LAYER PARAMETER D E P T H ,P-,S-VELOCITY,DENSITY (ISS(8) MUST=01

10 =0 NO PHASE-VELOCITY WINDOW= 1 ,2 , . . TWO-SIDED (COS**ISS(10))-WTNDOW

11 =0 NO FREQUENCY WINDOW= 1 ,2 , . . TWO-SIDED (COS**ISS(11))-WINDOW<0 COS**(1/ABS*ISS(11))-WINDOW

12 =1 EARTH-FLATTENING APPROXIMATION

IS APPLIED (ISS(8) MUST =1)

13 F R E E

14 = 1 THE CRITICAL DISTANCIES FOR ALL LAYERS ARE

CALCULATED AND WRITTEN ON UNIT 6

15 =0 LIST OF SEISMOGRAM IS PRINTED= 1 NO LIST OF SEISMOGRAMS

NO DIRECT WAVE IF MDECK.GE.ISO

EXPLOSIVE SOURCEDOUBLE COUPLEPOINT-SOURCE RADIATING P-,SV- AND SH-ENERGY,

16

17

= 1

=0=1- 1

P-INDEPENDENT=4 LINE SOURCE=5 SINGLE-FORCE

18 =0 FUCHS-MUELLER-SIGNAL= 1 DELTA-IMPULS=2 HEAVISIDE STEP-MPULS=3 MOMENTFUNCTTON AFTER BRUESTLE=4 NO SOURCE-FUNCTION - SPIKE, RECOMMENDED

FOR SPECTRAL ANALYSIS=5 DIGITIZED SOURCE SIGNAL, EXTRACTED E.G.

FROM REAL DATA (SEISMOGRAMS FOR VELOCITY)TO GET THE SAME SOURCE SIGNAL AS OUTPUTONE MUST CALCULATE THE SEISMOGRAMS FORVELOCITYAT THE END OF THE DATASET THE SOURCE SIGNALIS READ IN

=6 RICKER WAVELETTO GET THE SAME SOURCE SIGNAL AS OUTPUTONE MUST CALCULATE THE SEISMOGRAMS FORVELOCITY

=7 SOURCE-FUNCTION FOR SINGLE-FORCE

19 - ONLY IF(ISS(17).EQ.l)=0 READ IN THE MOMENT TENSOR=1 MOMENT-TENSOR AFTER AKI=2 MOMENT-TENSOR AFTER MUELLER(F1,..,N1,...)

20 - ONLY IF(ISS(17).NE.l)=0 INDEX OF THE SOURCE- AND RECEIVER-LAYER= 1 DEPTH OF THE SOURCE AND THE RECEIVER

21 - =0 NO MULTIPLE REFLECTIONS BETWEEN SOURCE ANDRECEIVER

= 1 MULTIPLE REFLECTIONS BETWEEN SOURCE ANDRECEIVER ARE CONSIDERED IF MDECK LE IREAND MDECK LE ISO

22

23

24

25

F

F

F

F

R

R

R

R

E

E

E

E

E

E

E

E

LINE 3: SOURCE-,REC7EIVER-LAYER, NO. OF LAYERS WITH TRANSMISSION ONLY,DENSITY TRANSFORMATION AND NO. OF SOURCES

ISO IRE MDECK NRHO NHFORMAT: 415

I S O J R E - IF ISS(17).NE.1.AND.ISS(2O).EQ.O , ISO.GE.l, IRE.GE.O

MDECK - NUMBER OF LAYERS ABOVE THE REFLECTIVITY ZONE,MDECK = 0 - FULL RESPONSE

NRHO - EXPONENT OF DENSITY TRANSORMATION IN CASE OFUSING THE EARTH FLATTENING-APPROXIMATION

NH - I F I S S ( 1 7 ) E Q . l - NUMBER OF SOURCES

LINE 3A: IF ISS(20).EQ.l AND ISS(17)NE.lDEPTH OF THE SOURCE DSO AND THE RECEIVER DREFORMAT: 2F10.4

LINE GROUP 4: MODEL-PARAMETER

THICKNESS D, P-VELOCITY, QP, S-VELOCITY, QS, DENSITYFORMAT: 6F10.4

OR DEPTH Z, P-VELOCITY, QP, S-VELOCITY, QS, DENSITY, NHSFORMAT: 6F10.4J10NHS = NUMBER OF LAYERS BETWEEN TWO ADJACENT Z, V VALUES

= 0 FIRST ORDER DISCONTINUITYIF S-VELOCITY NEG. VS=0. - LIQUID UPPERMOST LAYER (SEA)

LINE GROUP 4A: IF ISS(17).EQ.l OR ISS(17).EQ.5

1. DEPTH OF THE RECEIVER DREFORMAT: F10.4

2. SOURCE-COORDINATES (XS(I),YS(I),ZS(I),TS(I),ES(I),I=1,NH)TS - ORIGIN TIME FOR SOURCE NO. IES - STRENGTH OF SOURCE NO. IFORMAT: 6F10.4

3. ORIENTATION OF DOUBLE COUBLE OR SINGLE FORCEISS(19)=0: READ TENSOR OF MOMENTS: M11,M12,M13,M22,M23,M33ISS(19)=1: READ STRIKE,DIP,SLIP: PHIS,DELTA,LAMDAISS(19)=2: READ UNITVECTORS: F1,F2,F3,N1,N2,N3 -

PERPENDICULAR TO THE NODAL PLANESSINGLE-FORCE: READ FORCES: F1,F2,F3FORMAT: 6F10.4

LINE 5: SPECIFICATION OF DISTANCES, ANGLES AND NO. OF SEISMOGRAMSFIRST DIST. LAST DIST. DIST. INCREMENT, AZIMUTH, NO. OF DIST.

XI x2 DX AZI NENTFORMAT: 4F10.4J10

LINE 5A: IF NENT.NE.O: READ DISTANCES X(KE),KE=1,NENTFORMAT: 8F10.4

LINE 6: VRED - RED. VELOCITYTMIN - MINIMUM TIME OF SEISMOGRAM

VRED,TMINFORMAT: 2F10.4

L I N E 7 . C 2 - MINIMUM PHASE-VELOC, CWIL - LEFT WINDOWCWIR - RIGHT WINDOW, Cl - MAXIMUM PHASE-VELOCITYNP - NO. OF RAYPARAMETERSC2,CWIL,CWIR,C1,NPFORMAT: 4F10.4,110

L I N E 8 : F U - LEFT CORNER FREQUENCY, FWIL - LEFT WINDOWFWIR - RIGHT WINDOW, FO - RIGHT CORNER FEQUENCYFR - REFERENCE FREQUENCY FOR ATTENUATIONF R = 0 MEANS FREQUENCY INDEPENDENT ATTENUATIONIN CASE OF DOUBLE-COUPLE SOURCE FR IS SET TO ZEROFU, FWIL, FWIR, FO, FRFORMAT: 5F10.4

L I N E 9 : D T - TIME INCREMENT, NPTS - NO. OF VALUES,NA - NO. OF ZEROS IN FRONT OF THE SIGNAL,N - NO. OF EXTREMA IN THE SIGNALT - DURATION OF SIGNAL,TSIGMA (ANTIALIASING-FILTER, IF = 0 THEN NO FILTER)TSIGMA SHOULD BE 0.2-0.5*DT*NPTSTTBERG - HEAVYSIDE-LENGTH FOR SINGLE-FORCE WAVELETFOBERG - CONSTANT FOR SINLGE-FORCE WAVELETIF SPECTRALANALYSIS IS WANTED, TSIGMA MUST BE ZERODT, NPTS, NA, N, T, TSIGMA,TTBERG,FOBERGFORMAT: F10.4,3I10,4F10.4

LINE 10: ONLY IF ISS(18).EQ.5 DIGITIZED SOURCE SIGNALFORMAT: 8F10.4, END OF CARD SET BY INPUT OF -9999.

USED FILES: FOR SH : 3FOR P-SV: 2

- SEISMOGRAMS (FORMATTET)- SEISMOGRAMS (FORMATTET)

FOR SPECTRAL-ANALYSIS ONLY:21 - VERTICAL DISPLACEMENT(OMEGA)-FORMATTET22 - RADIAL DISPLACEMENT(OMEGA)-FORMATTET23 - TRANSVERSAL DISPLACEMENT(OMEGA)-FORMATTET

OUTPUTFORMAT FOR SEISMOGRAMS (Unit 2 or 3):

HEADERWORD - 7A6NLAY,MDECK,ISO,ISS4 - 415Z(I),D(I),A(I),B(I),RHO(I) - 5F10.4 I=1,NLAYNENT - NO. OF DISTANCIES - 15

10

VRED,TMIN,DT

- 7F10.3 I=1,NENT- 3F10.4

TRACE

R(KE),ABSAZ,IKOMP,NPTS,BALMAXISEIS(I), I=1,NPTS

- 2F15.5 ,15 ,- / I 1 0 , E 1 5 . 4- 1615

Note:

HEADER:HEADERWORDNLAY - Number of LayersMDECK - N u m b e r of t r a n s m i t t i n g l aye r sISO - Index of source layerISS4 - 0 - both vertical and radial component

1 - only vertical component2 - only radial component3 - transverse component

Z(I) - depth of layerD(I) - thickness of layerA(I) - P-Velocity of layerB(I) - S-Velocity of layerRHO(I)- Density of layerNENT - Number of distanciesX(I) - distanciesVRED - reduction velocity of calculated seismogramsTMIN - minimal time of calculated seismogramsDT - time increment

TRACE:KE - Loop over d i s t anc iesR(KE) - distance for actual traceABSAZ - starting time of actual trace in sekIKOMP - component (1 - vertical, 2 - radial)NPTS - Number of Points per traceBALMAX- True Amplitude Value for each traceISEIS - normalized Integer-Output, Zero-Line is 50000,

max. pos.Value is 99999, max. neg. Value is 1

To get-unnormalized traces you must do the following transformation:

DO 100 I = 1,NPTSSEIS(I) = ((ISEIS(I)-50000)/49999) BALMAX

11

OUTPUTFORMAT FOR SPECTRALANALYSIS:HEADER - 7A6NLAY,MDECK,ISO,ISS4 - 415

NENT - NO. OF DISTANCIES

VRED,TMIN,DTFU,FWIL,FWIR,FODNUECU,CWIL,CWIR,CONPR(KE),NPTS,SMAXS(JF) - COMPLEX

5F10.4 I=1,NLAY• 1 5

• 7F10.3 I=1,NENT• 3F10.4. //4F10.4• /F10.6• /4F10.4• / H O• /F10.4,I10,E15.5• 10F8.5 JF=l,NPTS/2

*******************************************************************

12

APPENDIX II, Data Analysis

Site and Data Description

The K-901 area chosen for the surface-wave tomography experiment is located at the Oak RidgeK-25 Plant on the Oak Ridge Reservation (Figure 3). The test site is approximately rectangularwith sides of about 750 by 450 ft. with mostly open or slightly wooded ground. In much of the areathe topography is smooth, although there is some significant elevation changes around some of theedges. Shots were placed on three sides, East, North, and South lines, all with recording on arraysof sensors on the opposite side. The spacing of the shots and sensors was 5 m. A 1500-lb airgun

Figure 3: Map of the K-901 test site. Shot and recording points are marked by solid dots. North istoward the right.

was used as the seismic energy source. The sensors were predominantly 10-Hz vertical geophonessuggesting sensitivity to frequencies as low as 5 Hz. The data were sampled at 1000 Hz, and recordedfor 8 s. The surface waves arrivals are in a time window of 0.5 to 2 s, suggesting average velocitiesof 100 to 300 m/s. The sample shot gather shown in Figure 4 was recorded along the west line forthe shot point 1001 located on the northeast corner.

Initial Processing

The original copy of the K-901 data were not readable due to a problem that must have occuredduring conversion from the SEG2 to SEG-Y format. The problem was resolved when the secondcopy of the data downloaded from the public FTP site at ORNL. Then, the next step was to editthe headers in data to correct shot and receiver locations. To save storage and CPU time, the datalength was reduced by half because there was no significant seismic information above 4 s. Finally,the original files are merged and saved in SU format, which is identical to SEG-Y format except forthe unassigned portion of the header is zeroed out.

Group Velocity Picking

In order to automatically estimate the group velocities from the travel time of a wave train at aparticular frequency, the multiple-filtering approach was employed. The existing program sugabor /included in the SU distribution was noted to be jagged amplitude sections and thus considered to

13

Shot #1001

40

3fc

CD

oQ)

cc20-

10-

.••.•/v.'.'-'.-v^-.' I l - \ r"

2Time [s]

Figure 4: Sample record section recorded along the west line.

be unstaisfactory. It is modified for the computation of instantaneous amplitudes by creating theanalytical signal in frequency domain. This new version is called sugaborl. The script mul t i f i l t erlisted below was used to display the filtered traces, multif i l ter requires two input parameters: shotnumber and center frequency of the narrow-band filter.

# ! /bin/shfname=$HOME/ornl/data/su/1001-1099.sushot=$lfcent=$3f las t=$[ fcen t+l lsuwind key=fldr min=$shot max=$shot <$fname | \sugaborl beta=0.001 band=l fmin=$fcent fmax=$flast | \suxwigb

The section in Figure 5 is an example of instantaneous amplitude sections computed for shot 1004of the K-901 data set. The section is obtained by issuing the following command:

% multifilter 1004 15

14

Figure 5: Instantaneous amplitudes at 15 Hz for shot gather 1004.

The main difficulty with this approach is that the arrival of spectral energy at the frequency ofinterest may not always be clearly identified as a sharp peak. On some of the traces the spectrumexhibits spectral splitting or superious peaks. This phenomenon is demonstrated in Figure 5. Notethat the arrivals on both sides of the section are marked by well defined spactral shape whilethere is a significant scatter in the middle. This scatter (or noise) in group arrival times will mostlikely degrade the resolution of the inversion scheme. When data length is limited, group velocityestimation with multiple filtering is known to be problematic in itself because of the inherent trade-offbetween spectral and temporal resolutions. The uncertainty introduced by the scatter in group traveltime picking is certainly an additional difficulty in implementation of the method. The automaticpicking of travel times is done by the program p ick-rayle igh which also prepares the input file forthe tomographic inversion program gt inv. This is achieved by piping the output of s u g a b o r l tosumaxl which finds the sample index of the maximum instantaneous amplitude.

#! /b in / sh

15

fname=$HOME/ornl/data/su/1001- 1099.sushotmin=$lshotmax=$2fcent=$3flast=$[fcent+l]outfile=$shotmin-$shotmax-$f cent "Hz ".picks

suwind key=fldr min=$shotmin max=$shotmax <$fname I \sugaborl beta=0.001 band=l fmin=$fcent fmax=$flast I \sumaxnew verbose=l mode=max > $outfile

pick-rayleigh takes three input parameters: initial and final shot numbers, and center frequeny.Between the frequency range of 10 to 24 Hz, group velocities are estimated with 2 Hz increamentsresulting in 8 sets of group velocities (Figure 6). For all 8 frequencies chosen, we observed a scatterin group velocity picks. The main feature of the disperson curve is that there is a gradual increasein group velocities up to 16 Hz and decrease toward 24 Hz.

P-wave travel time picking

The shell script pick-p was written to allow picking of first breaks.

#! /bin/shfname=$HOME/ornl/data/su/1001-1099.sushot=$lout file = $shot.pickssuwind <$fname key=fldr min=$shot max=$shotl \suwind tmax=0.12 l\suxwigb perc=40 cmap=hue mpicks=$outfile

When pick-p is invoked with the desired shot number (e.g. 1029), it displays on the screen the first120 ms of the data. Then by clicking the middle mouse button (both buttons pressed to emulate themiddle button) the x and y coordinates are obtained. To save the current reading into a file the Skey is pressed. The readings are saved in the file named shotl029.picks. Figure 8 shows the picksfor this example. Despite some variation with distance, the over all trend in P-wave travel-timeplot in Figure 9 indicates that the first breaks are purely refracted arrivals associated with a highvelocity (6500-7000 m/s) layer about 40 to 50 m deep.

Surface wave dispersion

The bedrock is cherty-dolostone of Cambro-Ordovician age belonging to the Knox group. The rockis very weathered and karstified. It is overlain by 10-60 ft of residual soils. Although the target of thetomography is areas between surface and about 100 ft, information on average velocity structure,near surface in particular, is of importance. For this purpose, we analyzed Rayleigh wave dispersionalong several paths (Figure 10).

The surface waves in the observed data exhibit inverse dispersion. Inverse dispersion is char-acterized by arrivals of the lower frequencies later than the high frequencies (see Figure 4). Ingeneral! inverse dispersion is an indication of a low velocity layer located below a high velocity layer.However, modeling of observed dispersion (Figure 11) suggest that the inverse dispersion is a resultof frequency content of the data that is controlled by the excitation of the energy source and theimpulse response of the geophones. Surface wave energy is predominantly between 10 and 25 Hzwith theoretical dispersion that is in a good aggreement with the observed data. For the frequencyrange of 10-25 Hz. the branch of the dispersion curve is clearly in the form of an inverse dispersionrelation. However, the theoretical layered model does not include a low velocity layer. Figure 12

16

Reflectivity synthetics

To verify the average velocity model derived for the site, a synthetic seismic section (Figure 13) isgenerated using the reflectivity method. The reflectivity method was developed for a horizontallyhomogeneous layered medium. The average (background) velocity structure is derived from thecombined results of the modeling of surface wave dispersion data and interpretation of the refractedP-wave arrivals associated with the deeper parts of the structure. The input file created for thereflectivity modeling is given below:

2-LAYER OVER HALF-SPACE0 0 0 11

1 10.00300.01400.023020.00000.00000.00000.00.1000

99999.0000.2100.0

0.0020

0 0 0 0 10 10.45000.78001.60006.6000

0.00000.00.2500.00.2401.002048

0 0 011000.01000.01000.01000.0

0.00001.00.005

6.040.0

0 1 0

0.2110.3570.7853.840

0.00000.0

7.00050.000

5 4 0 0 1

1000.1000.1000.1000.

1.00000.0 0

10000.000

1 0.23 1

0 0

2.52.52.52.5

.0

0.

Reflectivity synthetics are generated for a vertical point force applied at the free surface withan impulsive source time history function. Then, the output of the reflectivity run (the Green'sfunctions) are convolved with a Ricker wavelet whose center frequency is at 11 Hz corresponding toa maximum frequency of 22 Hz. This is approximately the highest frequency of energy existed foundin observed data. The distance range used in the computation of the synthetic traces was from 50 to250 m. The synthetic section is in good aggreement with the observed data (e.g. 4). To further testthe average-velocity model, we analyzed the surface wave dispersion on a synthetic seismic section.Dispersion estimated from the synthetic waveforms matches the theoretical results (Figure 14). Also,note that arrival of the higher mode is more prominent in the synthetics than the observed data.Given these constraints on the interpretation, a preliminary Single Value Decomposition inversionwas generated for frequencies from 10 to 24 Hertz. An example for 10 Hertz is shown in Figure 15.

17

Q3

• 3

QrOLf Vrtodty (frVi)

8 3 B 8 g IOnxpV«toctTy{n'*)

8 8 8 8 8

'Mm

8 8

;»«> .ko° o 8 o ,ft V

= »fsO 0

o ooo

P ^ ° o

43000

42900 -

42800 -

42700 -

42600 -

42500 -

42400 -

42300 -

42200 -

42100 -

42000

_1 I I I I !_

10400 10500 10600 10700 10800 10900 11000 11100 11200 11300 11400 11500 11600

Figure 7: Surface wave lines covering the K-901 site.

Sho11029

10

20

30

40

' 50

60 h

70

80

90

100

;

A

.-- r . ; •

1•"•>

mmm

~T 1

* v<> „&*<

2' "~ %

- :i ' ^

•T^jjia-* "'•'•'• - • £fyrLiffli*l*'';JMr

^ • i r ' ^ r - j S . ^ 2 *;.•••• i saft-*

—i 1—

m$ •**to

.« ' s _ * *

§&^^^

I I ^ ^ B I H mi mi TMI

1

s. ^ *I- " " •

»*«?- ,„

": ¥•i

I10 15 20 25 30

Recording Point35 40 45

Figure 8: P-wave picks for shotl029 recorded along the west line. Solid line is interpolated from thediscrete points picked by pick-p.

19

Shots 1001-1099

300

Figure 9: P-wave tra.vel time readings for the entire data set.

West

SouthFig. 12

Fig. 11top

Fig. 111middle i

North

East

Figure 10: Surface wave lines selected for the preliminary modeling.

20

Input File: shoti 002-trace30.d Input File: shoti 003-trace30.d

10 20 30Frequency (Hz)

0.80

0.70 -

0.60 i

0.50^

0.40^

0.30 H

0.20-i

0.10 j

10 20 30Frequency (Hz)

40

Input File: shot 1 002-trace48.d

10 20 30Frequency (Hz)

40

P - & S-Wave Velocity [ k m / s ]0.5 1

Figure 11: Group velocities measured in the middle of the west line for shot point located on thenortheast corner (top). Group velocity measurement on the northwest corner for the same shot pointas before (middle). Theoretical dispersion is indicated by solid lines. Layered model from inversion(bottom).

21

toto

Hn

P*vere

o-mod

n_

oa*

B.B'nO-

o3r-f-

crn

B'

i—*

L043

Ocn

oo3

. v

;do:

tn

to

Jnox

Q

o"

&-55'

o '

3

Group Velocity (km/s)

O O &.

ft to

2 e- ff-Op,"

o5 o

lie-

cr ri-

C3

0.01Depth [kml

5X1O"13

< • 1 1 1 1 1

-

oft?

CD

oo

Ol M

Group Velocity (km/s)

Group Velocity (km/s)o p o o o o o po o o o o o o o

Group Velocity (km/s)p p p p p p p p^ u Li * ai bi NJ t»o o o o o o o o

...I I I...

Figure 13: Synthetic reflectivity record section computed from the average velocity model for theK-901 site.

23

Input File: synt-trace31.dat

o0.10

10 20 30 40Frequency (Hz)

Figure 14: Dispersion relation estimated from 31th trace of the synthetic section.

24

kortk

Figure 15. Singular Value Decomposition inversion for 10, 14, 18, and 22 Hertz.

APPENDIX III. SEISMIC DATA ACQUISITION SYSTEMSYSTEM DESCRIPTION

System Overview

This document describes a data acquisition system for collecting and storing 16-bit analogdata from an array of seismic sensors following the activation of an impulse source (suchas a hammer). The hammer is fitted with an inertial switch that is connected to the triggerinput of an analog-to-digital converter (ADC) in a personal computer (PC). This datacollection system is built around a 16 channel Keithly Metrabyte ADC board which isplaced in a 486 PC computer. Analog data are sampled at an overall sampling rate of 98kiloHertz to provide a per channel sample rate of 6.125 kiloHertz. The voltage inputrange of the ADC is software controllable from a range of 10 volts to 1.25 volts full scale.The data collection software sets the input voltage range of the ADC to a bipolar mode ofplus 1.25 volts to minus 1.25 volts. This range is compatible with the output of the 16-channel preamplifier and low-pass filter unit.

This system consists of a hardware interface box consisting of 16 channels of preamplifierswith interconnections to the seismic sensors and to the ADC board in the computer(illustrated in Figure 1). The 16 seismic transducers can be arranged in any array patterndesired. Data collection is initiated by starting the data recording program designatingthe file name to be used for data storage, and striking the hammer on the ground at thedesired input point. Data collection starts at the close of the inertial switch attached to thehammer. The analog data lines and the trigger input are passed between the preamplifierinterface box and the ADC in the computer through a common ribbon cable.

The data collection program causes the ADC board to convert (with 16-bit resolution) andstore a total of 74,400 samples at a 6.25 kHz per channel sampling rate. This data recordrepresents a total time period of approximately 0.75 seconds. It is stored as an ASCII filewith each data point being followed by a carriage return and line feed. Data are recordedfrom each channel in the order of the channels from Channel 0 through Channel 15.

The data record is presented in a condensed graphical format on the computer monitor toallow the operator to make a quick "quality check" before proceeding to collect furtherdata records. The data collection routines are designed to allow the operator toautomatically cycle through a total of 5 successive data collection runs using the samegeometry. These files are recorded using the same file name with the file numberautomatically incremented with each run The operator may easily abort the multiple datafile collection by keyboard entry for single file recordings..

HARDWARE DESCRIPTION

The computer used with this system is a general purpose 486 computer with a hard diskcapacity greater than 2 gigabytes. The hard disk is used as the primary storage mediumfor data collection. High speed data collection is supported by direct memory access fromthe ADC during actual data conversion and collection. Data is transferred under programcontrol to the hard disk at the close of each data collection run

SEISMICSENSORARRAY

SWITCH

16-CHANNELPREAMPLIFIER

LOW-PASS FILTERUNIT

ANALOGDATALINES

\

486 COMPUTER

WITH

KETTHLEY METRABYTE

DAS-1BQ2HR ADC HOARD

EXTERNAL TRIGGERINPUT TO ADC CARD

IMPACT

Figure 1. Block Diagram of Seismic Data Collection System.

The ADC used in this system is the Keithley Metrabyte DAS-1802HR analog converterboard. This board is compatible with the 486/5 86/Pentium computer systems and acceptsup to 16 single-ended input channels or 8 differential input channels. The ADC boardrequires a full length slot in the computer. User connections to the board are madethrough a 50 pin I/O connector at the rear panel of the computer. All features of theboard are selected through the software except the base address of the board. Thesefeatures include unipolar and bipolar operation at four different input voltage ranges.

The base address of the DAS-1802HR ADC board is set by a 6-position DIP switchlocated on the ADC board. The setting used for the computer configuration included inthis data collection system is H310 hex. DMA channel 5 and interrupt level 10 are used inthis configuration. The configurations options for the ADC board are recorded in aconfiguration file that is used by the data collection software to properly configure theboard. This file should be in the same subdirectory as the executable data collectionprogram. This configuration file may be created or modified by using the CFG1800.EXEprogram supplied with the DAS-1802 support software. The menu options of thisprogram allow choice of (1) board type, (2) Base Address, (3) A/D mode, (4) A/Dcommon-mode ground reference, (5) DMA channel, (6) TRQ level, (8) number of SSHs(accessory boards attached), (9) SSH gains. (10) number of EXPs (total number of EXP-1800 accessories), and (11) number of ND channels. For further information on thisprogram see Section 3 of the Keithley Metrabyte User's Manual for the DAS-1800 boardseries. The configuration file used in the data collection system described here is listedbelow.

CFG 1800.

BoardNameAddressDMAChannelIntLevelADChanModeADChanConfig

FIL

0DAS1802HR&H310

510

BipolarSingle-ended

ADCommonMode LL-GNDNumSSHNumEXPADChannels

0016

PREAMPLIFIER/LO W-PASS FILTER CIRCUIT

The output of the seismic transducers must be amplified by a factor of approximately 50 tomake optimum use of the plus or minus 1.25 voltage operating range of the ADC asprogrammed in the supporting data collection software. The preamplifiers also contain anactive low-pass anti-aliasing filter to minimize interaction of frequencies in the convertedsignal samples with the sampling frequency. A diagram of the basic preamplifier circuit isgiven in Figure 2.

13.3 kOhmy—:wwwTransducer

Input

Ci1600 pfd

R36.8 kOhm

www-

c21600

68 kOhm

Output ToADC Card

Type 747 Operational Amplifier

Figure 2. Schematic diagram of sensor preamplifier and low-pass filter.

All preamplifier channels are identical. Each is composed of two cascaded operationalamplifiers operating in a single ended mode. The low frequency gain of each amplifier isset by the ratio of the feedback resistor to the input resistor. A gain of approximately 5 isobtained in the first amplifier and a gain of 10 is obtained in the second amplifier. Thecombination provides the desired overall preamplifier gain of 50. Eight preamplifiers arelocated on a single printed circuit card. Two cards are included in the 16 channel interfaceunit. The cards are illustrated in Figure 3. Connections to the cards are by ribbon cable.

POSITIVE SUPPLYVOLTA

NEGATIVESUPPLY

POWERGROUND

' • \ rx** •',«•'<•. ' *

NEGATIVESUFPLV

VOLTAGE

POSITIVESUPPLY

VOLTAGE

Figure 3. Eight channel preamplifier circuit board.

The low-pass filter is provided by the feedback capacitors around each operationalamplifier. The filters act independently due to the isolation produced by the amplifiers.The cutoff frequency is controlled by the product of the feedback capacitor and thefeedback resistor. The two operational amplifiers making up each preamplifier arecontained in the same integrated circuit chip. A plot of the low-pass filter characteristicsis included in Appendix D.

The preamplifier circuit cards are mounted on one side of the preamplifier/low-pass filterbox as shown in Figure 4. The input and output lines to these boards are ribbon cablesthat are permanently attached to the cards. The location of the two 6-volt lantern batteriesneeded to power the preamplifers can also be seen in this figure. The cable connectionsto these batteries is shown in Figure 5. The power connections are color coded: RED forpositive 6 volts, BLUE for negative 6 volts, and BLACK for ground. The batteries areheld in the compartment by a rotating bar over the top of the batteries. To remove thebatteries, press down on the top of the batteries and rotate the holder bar to clear thebattery compartment.

Figure 4. Locations of preamplifier cards and batteries.

Figure 5. Battery holders and battery cable connections.

connections between the preamplifier boards and the cable to the ADC board are madethrough the Preamplifier Interface Connector Board Located on the bottom of the toppanel. This board accepts the input and output cables from Preamplifier Board A andPreamplifier Board B (see Figures 6 and 7). The power cable from the batteries is alsoconnected to the rear of this panel where the power on-off switch is located.

Figure 6. Preamplifier interface connectors and cables.

Figure 7. Cable connections to rear of front panel.

The assembled preamplifier/low-pass filter box is shown in Figure 8. The power on-offswitch is located in the lower-right section of the panel. This panel also contains threedifferent connector groups. A fifty pin connector provides the interface connection to theADC board in the computer. Barrier terminal strips provide screw terminal connections to16 seismic sensors (see Figure 8). The order of the connections are indicated in thisfigure. Each sensor connection has two screw terminals. One is for the signal input andthe other is for the ground input. The sensor cable should be shielded for noise rejection.The shield of the sensor cable should NOT be connected to ground at the sensor end. Theshield should only be connected to the ground terminal at the interface box.

Figure 8. Front panel of seismic data acquisition interface box.

8.8 KILOHMS +6 VoltSupply

TRIGGERSWITCH

| NO

Pin 48

6.8 KILOHMS

Figure 9. Trigger circuit diagram.

The impulse trigger switch circuit is illustrated in Figure 9. Triggering occurs when TheTGIN input (pin 46) of the ADC connector goes positive. The trigger function is rearmedafter the completion of data collection and transfer of the data from the active memory tothe hard disk.

The 50 pin connector provides all interconnections between the ADC card in the computerand the preamplifiers. These connections include the 16 analog output signals from thepreamplifiers, the low-level signal ground to the amplifier box. and the trigger inputTGIN. A nodal ground is established in the interface box on the back of the top panel.The pin-out connections of this connector are defined in Figure 10.

(User-Common Mode)U_CMMD-01

CHOO LO or CH08 HI - 02CH01 LO or CH09 HI - 03CH02LOorCH10HI -04CH03LOorCH11 HI -05CH04LOorCH12HI -06CH05LOorCH13HI -07CH06LOorCH14HI -08CH07LOorCH15HI -09

ODAC22 - 10ODAC32- 11

+ 1 5 V - 1 2±15 V Return - 13

D GND- 14Dl 1 - 15D I 3 - 1 6

DO 1 -17DO 3 - 18

DOSTB- 19TGOUT - 20MUX 03-21MUX 05 - 22MUX 07 - 23

+5V - 24DGND-25

26 - CHOO HI27 - CH01 HI28 - CH02 HI29 - CH03 HI30 - CH04 HI31 - CH05 HI32 - CH06 HI33 - CH07 HI34 - LL GND35 - ODAC0'36 - ODAC1'37 - -15V38-±15 VReturr39 - GEXT40 - Dl 041 - D I 242 - DO 043 - DO 244 - XPCLK45 - SSHO46 - TGIN47 - MUX 0448- MUX 0649 - +5 V50 - D GND

Figure 10. Pin connections for ADC connector.

SOFTWARE DESCRIPTION

The supporting software consists of three basic program groups. These are (1) the bootdisk, (2) the data collection program, and (3) the data review or replay program. Theseprograms operate in the MS-DOS operating system and were written in basic to makemaximum use of existing software sub-routines. Each program was compiled to be astand alone executable programs using the Quick Basic 45 compiler.

The DOS Boot Disk programs should be operated from the A: drive. The data collectionprogram and the replay program are assumed to reside in a directory entitled "DAS 1800"on the C: drive of the computer. Data files are assumed to be stored in a sub-directory of"DAS 1800" entitled "DATA". The data collection and replay software also allows accessto data files located in any disk drive, directory, or sub-directory by entering the desiredlocation instead of accepting the default conditions. These programs may also beexecuted from a floppy disk drive if desired.

Boot Disk

The computer should be booted using the DOS Boot Disk to configure the computer in amanner that is optimum for the ADC and the supporting data collection program. ThisBoot Disk includes a DOS system the AUTOEXEC.BAT, and the CONFIG.SYS filesoptimized for data collection. The computer may be restored to its normal operating

configuration by removing the DOS Boot Disk and restarting the computer. Restartingcan be accomplished by turning the power of the computer off and back on (with about a5 second off delay) or by pressing CONTROL-ALTERNATE-DELETE simultaneously.The AUTOEXEC.BAT and CONFIG.SYS programs are listed below:

AUTOEXEC.BAT

C:\DOS\SMARTDRV.EXE@ECHO OFFPROMPT $p$gPATH C:\DOSSET TEMP=C:\DOSset mouse =c:\mousec:\mouse\mouse.exe /Qc:cd das1800

CONFIG.SYS

DEVICE=C:/DOS\SETVER.EXEDEVICE=C:\DOS\HIMEM.SYSDOS=HIGHDOS=C:\DOS\EEM386.EXE 1024FILES=30BUFFERS=20,0SHELL=C:\DOS\COMMAND .COM C:\DOS\ /p

The AUTOEXEC.BAT batch file on the DOS Boot Disk makes drive C: the active driveand selects the appropriate directory (DAS 1800).

SEISMIC DATA COLLECTION PROGRAM "SEISMIC.EXE"

This executable program collects 16 channels of seismic data using a 16-bit DAS-1800Metrabyte analog-to-digital (ADC) converter board. It operates at the maximum samplerate of the ADC (98 kHz) to give a per channel sample rate of 6.125 kHz. The softwareassumes 16 single ended channels operating in a bipolar mode. The gains for all channelsare set by the software to a full-scale value of 1.25 Volts. This program requires that theconfiguration program DAS1800.CFG be in the same sub-directory at the time ofexecution. A listing of this program is included as Appendix B of this document.

The data collection program is started at the DOS prompt C:> by entering "SEISMIC"and pressing "ENTER'. When the program loads, the operator is requested to enter a 6character file identification and press return as shown in Figure 11. These characters maybe any combination of upper and lower case letters and numbers.

DATA COLLECTION PROGRAM

FOR ACOUSTIC/SEISMIC IMAGING EXPERIMENTS

Collects 5 runs to same file name.

ENTER 6-DIGIT NAME FOR DATA STORAGE FILE

RUN NUMBER AND FILE SUFFIX WILL BE ADDED AUTOMATICALLY? tstfll

w * w * t w i » t * w * * w « « w w w t t ? t

DEFAULT DESTINATION DRIUE IS C:\DAS1800SDATAS

PRESS ENTER TO ACCEPT OR ENTER ALTERNATE DETINATION

(example - A:\Data\)

Figure 11. Input screen for selecting data file and location path for SEISMIC.EXE.

In the example in Figure 11, the first file will be labeled as tstfll01.DAT. Up to fouradditional data files may be recorded to this same 6 character file name if desired with thelast two digits of the file name being automatically incremented (tstfllO2.DAT,tstfllO3.DAT, etc.).

The second input request to the operator (see Figure 11) is for the location where the datawill be stored. The default location (C:\DAS1800\DATA\) may be accepted by pressing"ENTER." This assumes that the directory "DAS 1800" and its sub-directory "DATA"exists. If these are not present on the C: drive, then an error will occur and the programwill terminate. Other data storage locations (including different disk drives) may beselected by entering the location as shown in the example in the example line in Figure 11.

Figure 12 shows the display screen content following designation of the data file storagelocation. At this point, the ADC is armed to execute data recording on the occurrence ofa trigger signal from the switch attached to the impact source (such as a hammer).

Figure 13 shows the display during data collection and data recording. Data is stored inthe computer's RAM by direct memory access (DMA) under control of the ADC. Whendata acquisition is complete, the data is down-loaded to the designated file name byreading five different buffers. The progress of this transfer is shown in the lower rightcomer of the display screen.

DATA COLLECTION PROGRAM

FOR ACOUSTIC/SEISMIC IMAGING EXPERIMENTS

Collects 5 runs to sane file name.

ENTER 6-DIGIT NAME FOR DATA STORAGE FILE

RUN NUMBER AND FILE SUFFIX WILL BE ADDED AUTOMATICALLY? tstf11

DEFAULT DESTINATION DRIUE IS C:\DAS1800\DATA\

PRESS ENTER TO ACCEPT OR ENTER ALTERNATE DETINATION

DATA ACQUISITION STARTS ON NEXT EXTERNAL TRIGGER.

Status: 101 Count : 0Driver's Far Heap Size: 182912QB's Far Heap Size: 182912Press a key to STOP Acquisition...

Figure 12. Data collection program ready for data acquisition.

Status: 510Driver's FarQB's Far HeapPress a key to

DATA COLLECTION PROGRAM

FOR ACOUSTIC/SEISMIC IMAGING EXPERIMENTS

Collects 5 runs to sane file nane.

ENTER 6-DIGIT NAME FOR DATA STORAGE FILE

RUN NUMBER AND FILE SUFFIX UILL BE ADDED AUTOMATICALLY? tstf11

DEFAULT DESTINATION DRIUE IS C:XDAS1800\DATAS

PRESS ENTER TO ACCEPT OR ENTER ALTERNATE DETINATION

DATA ACQUISITION STARTS ON NEXT EXTERNAL TRIGGER.

Count : 14880 RECORDING DATA TO tstfU01.DATHeap Size: 182912 UNLOADING BUFFER 2Size: 18Z912STOP Acquisition... Data acquisition completed

Figure 13. Data Collection screen during recording.

After the data file is transferred to disk, the first 150 milliseconds of the file is presentedon the display as shown in Figure 14. This graph shows the input signals in order withchannel zero at the top. Increasing time is to the right. This plot allows the operator tocheck the quality of the data recording and to make sure that all sensors are functioning.If the data is not as expected, then the data can be overwritten by restarting the programusing the same file name. Recordings of the data plots generated by the SEISMICprogram can be made using a screen capture program. The capture program may requirethat it be loaded before starting the SEISMIC data collection.

This plot shows signals recorded from three seismic sensors during system checkout. Thesensors were connected to inputs 0, 3, and 6 and placed in a line on the floor of the lab atapproximately 2 meter spacing. Impact was nearest to channel 6 sensor and the impactwave front across the other two sensors is evident in the time delay of the plots.

The file name and location of the data file are shown at the top of the graph along with thedate and time of the recording. Additional data files may be recorded under the same 6-character file name by pressing R to clear the screen and rearm the ADC for the nexttrigger input. To exit the current data run and quit the program press Q.

SDAS1800\DATA\test0101.DAT Q=QUII. R=RUM NEXT FILE12-12-1996 15:51:16

Figure 14. Graphical presentation of 16-channel data file in SEISMIC.EXE program.

SEISMIC DATA REVIEW PROGRAM "REPLAY.EXE"

Data tiles may be reexamined at any time by using the program "REPLAY.EXE." Thisprogram is started by typing "REPLAY" and pressing "ENTER." The operator isquestioned (see Figure 15) to enter the file name and the location of the data file. In thisexample, an alternate file location has been entered. When the file name and location areentered, the first file in the sequence is presented on the display. Figure 16 shows thereplay plot of the same data file shown previously in Figure 14. The REPLAY programplots the entire data file so the time scale appears compressed when compared to the plotassociated with the SEISMIC program. Recordings of the data plots generated by thisreplay program can be made using a screen capture program. This may require that thescreen capture program be loaded into high memory in the computer prior to execution ofthe replay program. This type of utility usually stores the screen image to disk for laterrecovery and examination. A listing of this program is included as Appendix C of thisdocument.

The REPLAY program can be cycled to read the next file under the current file name (ifthe file number is less than 5) by pressing "ENTER". To enter a new file name press R.To quit the program press Q.

SEISMIC DATA REUIEU PROGRAM

For 16 Channel Data F i l e s

ENTER 6-DIGIT NAHE FOR DATA STORAGE FILE

? TSTFIL

DEFAULT DESTINATION DRIUE IS CADAS1800SDATAS

PRESS ENTER TO ACCEPT OR ENTER ALTERNATE DETINATION

(example - A:\DataS)? C:\QB45N|

Figure 15. Input screen for selecting data file and location path.

C:\DAS1800VDflTAVtest0101.DAT

Figure 16. Graphical presentation of 16-channel data file in replay program.

DATA FILE FORMAT

The data files generated by the data collection program and read by the replay program arein an ASCII format. Each data sample is an integer in the range between -32768 and+32767. Each entry in the file is followed by a carriage return and line feed. The firstentry in the file is from input channel zero. The remaining channels follow in numericalsequence. This sequence is repeated until the end of the data file. If necessary, the datafiles can be examined by using DOS TYPE "filename.DAT" command. The extension " Imore?' should follow the type command to allow the data to be read by paging through it.

Appendix IV Validation of Shallow Seismic Structure at the K-901 BurialGrounds, East Tennessee Technology Park.

Appendix IV is a statement of proposed field work and background studiescooperatively developed by Dr. William Doll, at Oak Ridge National Laboratory, and Dr.Leland Timothy Long at Georgia Tech. The field work will be coordinated for joint useof ORNL equipment and Georgia Tech equipment.

1. BackgroundIn 1995, surface wave tomography data were collected by Oak Ridge national

Laboratory and the Kansas Geological Survey in the vicinity of the K-901 BurialGrounds at the East Tennessee Technology Park (ETTP, formerly known as the OakRidge K-25 Plant). The data were acquired in association with acquisition of seismicreflection data in the vicinity, and with direction from Dr. L. T. Long of the Georgiainstitute of Technology with the intent of testing new techniques that were indevelopment at Georgia Tech. In fiscal year 1996, Dr. Long received funding from theDepartment of Energy's Environmental Management Science Program (EMSP) todevelop the method. Preliminary models for the K-901 site, based on the surface wavedispersion inversion have been developed by Drs. Long and Kocaoglu. Thesepreliminary models raised questions that can only be answered effectively by additionalfield measurements of seismic velocity and the velocity structure.

2. ObjectiveThe objective of the proposed field measurements and studies is to acquire

ancillary data that can be used to validate the preliminary results of the inversion ofsurface waves, and that can assist in further development of the technique.

3. Scope

Task 1: To review driller's logs and historical documents to determine whetherfill material has been placed above the natural geologic materials at the site, and if so, thelocation and thickness of the fill materials.

A number of well have been drilled in the vicinity of the site during an SAICinvestigation in 1994-1995. In addition, Phil Coleman and others at the ORNLGeographic Information Science and Technology group have prepared fill thickness mapsof the K-25 plant vicinity based on subtraction of old (ca. 1945) topographic maps fromcurrent maps. These materials, as well as any other relevant materials that becomeavailable, will be reviewed as part of this task.

Task 2: Processing of a portion of seismic reflection Line C adjacent to the site.

The seismic reflection data that were acquired in 1995 have not been processed beyounthe brute stack. This task will provide shallow structural information in the form of

refraction statics corrections, that will provide the depth to bedrock as well as thevelocity of the bedrock along the east side of the study area. In addition, perturbations inthe continuity or amplitude of shallow reflections will provide evidence for structuraldisturbances that extend into the near surface.

Task 3: Acquisition and processing of high resolution seismic refraction datawithin the study area.

High resolution reversed seismic refraction data will be acquired across two or more ofthe (low or high velocity) anomalies identified in the preliminary surface wave dispersioninversions from the study area. These data will be acquired with an accelerated weightdrop source along three parallel lines across each of the selected anomalies, and thenalong another three lines that are orthogonal to the first ones. In addition to shootingfrom the ends and center of these refraction lines, shots will be fired from points alongsome of the orthogonal lines and from points off the end of the lines. This geometry willprovide high resolution imaging of the bedrock surface beneath the anomalies, and willhelp o discriminate three dimensional structures from one- and two-dimensionalstructures. A 48-channel Geometries Strataview seismograph will be used in conjunctionwith 10 Hz geophones. The geopohone spacings and shot geometry will be designed toprovide the greatest detail of the anomaly and surrounding velocity structures.

Task 4: Acquisition of fan refraction data within the study area

A fan geometry will be used to acquire surface wave data at up to three locations that willbe specified by Georgia Tech, using the same energy source as indicated in Task 2(accelerated weight drop) and recorded with low-frequency (1 Hz) geophones. Thesedata will be acquired with a geometry that will be specified by Georgia Tech and will beprocessed by Georgia Tech. The low frequency geophones will be provided by GeorgiaTech, and adapters to mate them with the ORNL Cable will be purchased, if necessary.

Task 5: Acquisition and processing of vertical seismic profile data within thestudy area.

Vertical seismic profile measurements will be acquired in the vicinity of one or two of theanomalies using Innovative Transducers DF-5 hydrophones and/or a 3-componentdownhole geophone. A sledge hammer or accelerated weight drop source will be usedfor an energy source. The data will be acquired in a temporary PVC-cased well(s) thatextend(s) to bedrock and shotpoints will be placed in at least four profile lines extendingaway from the well.

4. Deliverables

Products form this work will include the following:

A. A written summary of the historical review of the site that will include a descriptionof previous geotechnical results at the site and information on the fill thickness andlocations at the site.

B. A processed Common Midpoint Stacked section of the seismic reflection data thatwere acquired at the site along with plots of the thickness of the surface layer andvelocity of bedrock that are determined when applying refraction statics analysis tothese data.

C. A written summary of the seismic refraction results including a description of themethods that were used, plots of first arrivals, and resultant models, as well as ananalysis of these results.

D. Data acquired with a fan geometry, as indicated earlier, and in SEG-2, SEG-Y, orASCII format, as specified by Georgia Tech.

E. A written summary of the veritcal seismic profile data acquisition, processing andresults.

10